Date Presented 04/06/19
Motion tracking sensors were added to a commercial SaeboGlove, enabling patients to engage in manual practice of instrumental activities of daily living in a computer-generated virtual world. Sensor-derived performance in completing functional tasks in the virtual environment was compared to real-world performance assessed using gold-standard measures. Statistical analysis of data collected during use by 15 stroke patients confirms the concurrent validity of the computer-generated measures.
Primary Author and Speaker: Allison Ellington
Additional Authors and Speakers: Richard Adams
Contributing Authors: Kate Armstead, Kristen Sheffield, Heather Standish
PURPOSE: Investigate the validity of metrics derived during virtual IADL practice involving a sensor-augmented SaeboGlove and SaeboVR software for assessment of upper extremity (UE) motor function. This research is an important step towards enabling remote tracking of patient performance during independent, home-based practice and teletherapy.
DESIGN: Participants in this experimental study were hemiparetic stroke survivors with ongoing hand function impairment, recruited from a population of current and former rehabilitation clients. Inclusion criteria included minimum levels of active finger and arm range of motion. The primary investigation hypothesis was that measures derived during practice of virtual activities are strongly correlated to accepted gold-standard measures of UE performance.
METHOD: Study sessions were conducted by practicing occupational therapists within an outpatient clinic. Following consent and glove fitting, participants were asked to return for four virtual IADL practice sessions. In each, the participants completed a range of virtual activities involving repetitive, occupation-based motor tasks, such as cooking, shopping or pet care. The specially-instrumented SaeboGlove and SaeboVR software (Adams et al., 2017) were used to capture joint angle histories and generate metrics related to the speed of movement (subtask completion time – SCT) and smoothness of movement (normalized speed – NS) (Adams et al., 2015). In a final session, therapists administered the Wolf Motor Function Test (WMFT), the Fugl-Meyer Upper Extremity (FMUE) assessment, the Box and Block (BAB) Test, and Motor Activity Log (MAL).
The bivariate relationships between each sensor-derived measure and each gold standard clinical measure were estimated by way of the Pearson Product-Moment (PM) correlation coefficient. For each correlation analysis a 95% confidence interface (CI) was formed. A p ≤ 0.05 decision rule was used to establish statistical significance.
RESULTS: A total of 17 stroke patients were consented, and 15 completed the protocol. A summary of study characteristics includes: gender: female – 7, male – 10; race category: black – 3, white – 14; median age – 67 years; and median time since stroke – 12 months.
A strong and statistically significant relationship was found between the sensor-derived SCT (speed of movement) measure and all the gold standard clinical measures. The Pearson PM correlations were as follows:
WMFT: r=0.94 (CI: [0.82 0.98], p<0.001);
FMUE: r=-0.70 (CI: [-0.89 -0.27], p=0.004);
BAB: r=-0.56 (CI: [-0.84 -0.05], p=0.030);
MAL: r=-0.55 (CI: [-0.83 -0.03], p=0.035);
A strong and statistically significant relationship was also found between the sensor-derived NS (smoothness of movement) measure and the WMFT score, r=-0.81 (CI: [-0.94 0.49], p<0.001). The correlation between NS and the other gold-standard measures was not significant.
CONCLUSION: The primary investigation hypothesis of this study is confirmed. A strong and statistically significant correlation was found between sensor-glove generated metrics derived during virtual task practice and widely-accepted gold standard measures. We thus find that a sensor-augmented SaeboGlove, used in conjunction SaeboVR, can support valid assessment of UE motor performance.
Future research will assess the efficacy of repetitive occupation-based task practice that involves the hand in virtual IADLs using the sensor-enabled SaeboGlove over a course of independent home-based practice by stroke patients.
IMPACT STATEMENT: By establishing the validity of computer-generated tracking measures in a system intended for independent, home-based practice, this research provides a foundation for new approaches to therapy delivery involving remote monitoring and/or teletherapy.
References
Adams, R. J., Lichter, M. D., Krepkovich, E. T., Ellington, A., White, M., and Diamond, P. T. (2015). Assessing Upper Extremity Motor Function in Practice of Virtual Activities of Daily Living. IEEE Transactions on Neural Systems and Rehabilitation Engineering, 23, 287-296. doi:10.1109/TNSRE.2014.2360149
Adams, R. J., Lichter, M. D., Ellington, A., White, M., Armstead, K., Patrie, J. T., and Diamond, P T. (2017). Virtual Activities of Daily Living for Recovery of Upper Extremity Motor Function. IEEE Transactions on Neural Systems and Rehabilitation Engineering, 26, 252-260. doi:10.1109/TNSRE.2017.2771272
